Getting Started¶
Installation¶
pip install perfusio
For GPU support:
pip install perfusio[gpu]
For the interactive dashboard:
pip install perfusio[dash]
From source (development)¶
git clone https://github.com/lynchaos/perfusio.git
cd perfusio
pip install -e ".[dev]"
Quick-start¶
from perfusio.simulator.cho_perfusion import CHOSimulator
from perfusio.viz.theme import apply_theme
from perfusio.viz.static import fig4_training_trajectories
apply_theme()
sim = CHOSimulator(clone="CloneX", seed=0)
runs = sim.generate_box_behnken_experiment(n_days=28, seed=0)
# generate_box_behnken_experiment returns 27 runs (4-factor BBD, 3 centre points)
# Pass clone_labels to split left/right subplots by clone identity;
# omit it to split evenly by run index.
fig = fig4_training_trajectories(runs)
fig.savefig("fig4.pdf")
Reproducing the paper figures¶
python examples/reproduce_paper_figures.py
All figures are written to paper_figures/ in PDF, PNG, and SVG format.
Running the CLI¶
# Generates 27 runs (4-factor Box-Behnken design, 3 centre points)
perfusio simulate --clone CloneX --n-days 28 --out runs/
perfusio train --runs runs/ --model-out model.pt
perfusio run --model model.pt --days 28 --clone CloneX --dashboard
perfusio reproduce-figures --data runs/ --out paper_figures/
Citation¶
If you use perfusio, please cite both the original methodology paper and this library:
@article{gadiyar2026,
author = {Gadiyar, Chiraag J. and others},
title = {Self-Driving Development of Perfusion Processes for Monoclonal Antibody Production},
journal = {Biotechnology and Bioengineering},
year = {2026},
volume = {123},
number = {2},
pages = {391--405},
}